mirror of
https://github.com/pbakaus/impeccable.git
synced 2026-09-12 22:26:38 +03:00
4f7fdeca6bde016e04cb3ef1f227e65ca2dee8ea
14
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
   
|
36457e191f |
Bump skill-behavior test lineup: gemini-3.7-flash (#598)
* Bump skill-behavior google lineup to gemini-3.7-flash gemini-3.7-flash replaces gemini-3.6-flash in DEFAULT_MODELS. The README notes that the recorded gemini baseline cells were measured on 3.6-flash (or 3.5-flash where marked) and count as unmeasured on 3.7 per the suite's own cross-version rule, to be re-run on the next Setup or routing change. AI-assisted change. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Potential fix for pull request finding Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> |
||
|
Paul BakausandClaude Opus 5
|
65de2d294b |
Raise the skill-behavior timeout that was grading haste over thoroughness
`initialized natural build` looked like a third defect on main: sonnet began implementation before the attended concept checkpoint, three runs in a row. It is flaky, not broken, and the measurement setup was the larger problem. A run that stops to put the concept to the user before building takes about 579s on sonnet. A run that skips the checkpoint and fails the assertion finishes in 130-200s. The suite capped each test at 300s, so the thorough path was killed as a timeout and the hasty path was graded as a result: the cap was selecting for the behavior the scenario exists to forbid. Raised to 900s, with the reasoning recorded next to the number so it is not trimmed back as a mystery constant. The baseline is corrected accordingly: the scenario is flaky (1 of 4), not failing, and readers are told to check a duration against the cap before calling a slow failure a behavioral one. Written with AI assistance (Claude Code). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
Paul BakausandClaude Opus 5
|
c489335799 |
Build path becomes a config key existing projects can actually reach
The build-path preference shipped as a question only `init` asks, written to a file only `init` writes. Nothing routes an initialized project back through init, so every existing project took the comp-first default without anyone choosing it, and the only recourse was a footer toggle that binds one session. Neither the setting nor the round that preceded it ever reached a release (skill-v4.0.4 has no `buildPath`, no `comp-led`, no `.impeccable/settings.json`), so the PRODUCT.md standing-commitment fallback describes an era that never existed publicly. It is deleted rather than honored: told a field might exist, models go hunting for it and preserve it. - `buildPath` moves from `.impeccable/settings.json` into the unified `.impeccable/config.json`, which already has a known-keys registry, doctor coverage, and a gitignored `config.local.json` override. Whether a machine has an image tool is a property of that machine, so the local file wins. - new-work captures the answer from behavior instead of an interview: a toggle flip on a project recording nothing asks once, after the round closes, whether to keep it. The answer is written either way, because a declined offer nothing writes down is an offer the next session makes again. - Two findings: `config-invalid-build-path` (an unread value rides the default rather than the opposite path) and `config-build-path-unset`, gated on a product record plus evidence of direction work so polish-and-audit projects never hear about a setting they do not use. - init treats a recorded value as a confirmed answer, resolving its conflict with Step 1's "do not reopen confirmed fields". - The setting was undocumented in the README and doctor.md. Both now cover it. Also records a measured skill-behavior baseline. Three cells fail on unmodified main (scenarios 9 and 15, `initialized natural build`), verified against a clean worktree; the suite README now says so, so the next person does not spend the hour attributing them to their own branch. Written with AI assistance (Claude Code). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
Paul BakausandClaude Opus 5
|
121602079c |
Deliver the report as its own step; retune the lineup
Two failures the trace test found were structural, not model quirks. critique.md described the report's format and then went straight to writing a temp file, with no step saying to output the report. gemini-3.6-flash and luna both responded by bundling heredoc, snapshot write, trend read, and cleanup into one bash call and stopping, leaving a perfect archive nobody had read. A "Deliver the Report" step now precedes persistence, and persistence describes itself as a copy of what was already sent. gemini-3.6-flash failed three consecutive runs before this and its failures afterward all show the report reaching chat. The close is also step 6 of the persistence list rather than a section after it, since the same shape is what fixed delivery. Lineup: gpt-5.6-luna and deepseek-v4-flash out, gpt-5.6-terra in at reasoningEffort high (IMPECCABLE_SKILL_BEHAVIOR_EFFORT overrides), gemini 3.5 to 3.6. Provider options resolve from the model object inside the harness so the 21 runTurn call sites are untouched. Verified the effort actually reaches the API rather than being silently dropped. The Gemini bump was not cosmetic: 3.5-flash passed critique closes twice and 3.6-flash then failed three times against identical text. A version bump inside one family changed the outcome, so the README now treats cross-version carryover as unmeasured. Known floor, recorded: critique closes is flaky on gemini-3.6-flash, 1 run in 3. Two structural attempts moved it from consistently failing to intermittently passing and then stopped paying. claude-sonnet-5 and gpt-5.6-terra are clean. Prepared with AI assistance (Claude Code). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
Paul BakausandClaude Opus 5
|
ebc63f071a |
Fix critique's close on the right mechanism
The earlier fix in this branch was built on a wrong diagnosis. It assumed a
structured question hides any prose sharing its message, so it split report and
question across two turns. A controlled check showed prose before a question
renders fine; what hides a report is emitting it AFTER the question. The split
therefore fixed nothing and introduced a worse failure: a turn that ends on the
report is a turn that ends, and the questions never arrived at all.
Persistence returns to main's ordering, byte for byte, and the boundary prose is
gone. What replaces it is a position rule: the question is the last thing in the
response.
The trace test added here found two failures beyond the reported one. Critique
can fail to land in three ways, and they are now all asserted:
1. Question emitted before the report, hiding it behind the picker.
2. No close at all: no questions and no skip line, so polish inherits nothing.
3. Report authored into the persistence heredoc and never written to chat,
leaving a perfect snapshot and a user who sees nothing.
Mode 3 predates this branch entirely. Persistence step 1 now says the temp file
is an archive copy, not delivery.
The Codex final-question gate is promoted out of its <codex> fence, where it was
stripped for three of four providers, and the skip branch is now a countable
threshold (fewer than 3 Priority Issues) rather than a judgment call.
Known floor, recorded in the suite README: gpt-5.6-luna passes 1 run in 6 and
deepseek-v4-flash is flaky. claude-sonnet-5 and gemini-3.5-flash are consistent.
Prepared with AI assistance (Claude Code).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
||
|
Paul BakausandClaude Opus 5
|
0e5c6cbe17 |
Keep critique's report out of the question's message
The critique report and the AskUserQuestion call shipped in one assistant message, so the report stayed hidden until the user answered the picker and the command read as if it had never run. Reorder critique's persistence steps so the temp-file cleanup runs after the report and trend line are sent. That cleanup now ends the message carrying the report, leaving the questions to open a fresh one. Both critique.md and overdrive.md state the constraint and why it exists, so the ordering is not an unexplained sequence a model can optimize away. Overdrive additionally moves its direction descriptions inside the question options, where the user is actually reading them. Also fix the ask_instruction splices. The placeholder is a complete sentence, but five call sites spliced it mid-sentence and shipped text like "stop and STOP and call the AskUserQuestion tool to clarify. before expanding it". Every call site is now sentence-initial and the twelve lowercase provider values are capitalized to match, with a comment in utils.js pinning the contract. Record a workflow-contract baseline for the current model lineup. The two failures seen while validating this change are pre-existing: bolder refinement fails on deepseek-v4-flash identically with bolder.md reverted to HEAD, and redesign replaces DESIGN is flaky on assertions driven by new-work.md, which this change does not touch. Prepared with AI assistance (Claude Code). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
Paul Bakaus
|
77c7d8e0fc | Refine product and visual work lifecycle | ||
|
Paul Bakaus
|
bbed6eef08 |
Refresh the Impeccable product experience
Rework the landing page proof, steering demo, feature grid, slop catalog, detector coverage, theming, Live workflow, and responsive behavior.\n\nAI-assisted implementation by OpenAI Codex. |
||
|
Paul Bakaus
|
ead6ddabe5 |
Preserve experimental Live app-server workstream
Snapshot the current app-server implementation, shared Live optimizations, generated harness output, and in-progress site work before restoring polling as the primary runtime path. Prepared with Codex assistance under maintainer direction. |
||
|
|
c11cc7b58c |
Route native projects to native command variants (audit, adapt) (#357)
* Route native projects to native command variants for audit and adapt Follow-up to #269. The web audit.md and adapt.md carried "translate this yourself" Platform notes, so a native invocation paid for the full web file (~1.8k / ~2.6k tokens, mostly inapplicable) and did error-prone run-time translation. Authored with AI assistance (Claude Code) under maintainer direction. - New reference/audit.native.md and reference/adapt.native.md: authored native content (VoiceOver/TalkBack, platform conformance, adaptivity dimensions; phone-to-tablet, platform-to-platform, web-to-native strategies). One variant per command covers ios, android, and adaptive; per-OS specifics stay in the platform refs Setup loads regardless. - SKILL.src.md: Commands table lists the variants; Setup step 2 reads the variant instead of the web file when the platform is native. - audit.md / adapt.md: Platform sections replaced with a one-line web-only guard pointing at the variant. - animate.md / layout.md: Platform sections deleted; the Motion and Layout sections of the already-loaded platform refs carry that content. Web users now pay zero tokens for the platform axis in these files. - Skill-behavior scenario 15 pins the route-instead behavior (passes live on claude-sonnet-4-6); CLAUDE.md documents the variant convention. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Phrase command-reference routing as one rule, not rule-plus-exception Copilot review catch: step 2 said "MUST read reference/<command>.md" and then carved out the native variant, which invites loading both files. Now a single rule: read the web reference or the table's native variant, one file, not both. Scenario 15 re-verified live. Applied with AI assistance (Claude Code) under maintainer direction. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Anchor native runs in animate/layout, drop loaded-refs assumption Review-thread fixes, applied with AI assistance (Claude Code) under maintainer direction: - Greptile: deleting the animate/layout Platform sections left native runs alone with web tooling instructions (CSS keyframes, GSAP, Grid, clamp()). Restore a one-line anchor in each pointing at the loaded platform reference's Motion / Layout section (~20 tokens, not the old restatements). - Bugbot: audit.native.md and adapt.native.md asserted the platform refs were "already loaded in Setup", but the command reference loads at step 2, before step 5. Now they instruct: read the platform reference first if Setup hasn't already. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Carry the native-variant rule into routing rules 2 and 3 Bugbot catch: Setup step 2 routed native projects to the variant, but routing rules 2 and 3 (the operative text at command time) still said to load the generic reference file. Both now reference the same one-file variant rule. Applied with AI assistance (Claude Code) under maintainer direction. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Point animate/layout native anchors at the files, not "loaded" refs Bugbot catch, same class as the variant wording fix: the anchor lines said "the loaded platform reference" but command files load at step 2, before the platform refs at step 5. Both anchors now name the files and instruct reading them first if Setup hasn't already. Applied with AI assistance (Claude Code) under maintainer direction. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com> |
||
|
3e38e595c7 |
Add platform axis (web / ios / android / adaptive) (#269)
* Add a platform axis (web / ios / android / adaptive) to the skill Orthogonal to register: register decides whether design IS or SERVES the product; platform decides the delivery target and which native conventions apply. Set `## Platform` in PRODUCT.md; a missing field defaults to `web`, so legacy projects are unaffected. - extractPlatform() in skill/scripts/context.mjs (mirrors extractRegister); the CLI appends a NEXT STEP directive to read the native reference(s). `adaptive` (Flutter / RN / KMP shipping both iOS and Android) loads both ios.md and android.md. - New reference/ios.md (Apple HIG distilled) and reference/android.md (Material 3 distilled); reference/web.md is a thin pointer. The native refs frame register's role as narrow: platform conformance is the bar, brand lives in the expressive layer the platform gives you, never by breaking the rails. - Setup step 5 loads the native reference(s) when platform is native. Live mode and the detect CLI stay web-only, gated off ios/android/adaptive. - init asks platform right after register; adapt/audit/animate/layout carry short platform divergence notes; all secondary spots thread `adaptive`. - a11y stays in audit.md (loading it at design time makes output timid), so the native refs carry no Accessibility section; audit.md's Platform section owns native a11y. - Tests: extractPlatform unit coverage + skill-behavior scenario 10 (PRODUCT.md platform ios -> agent loads ios.md). Source-first: only skill/, scripts/, tests/, CLAUDE.md, NOTICE.md, the changelog and version are committed; the sync workflow regenerates the provider trees and ./plugin on merge. ios.md / android.md are distilled from the MIT-licensed ehmo/platform-design-skills; attribution in NOTICE.md. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Address review: gate web tools on native platforms, drop version churn Maintainer-review fixes applied with AI assistance (Claude Code), on top of the rebased platform-axis commit: - Design hook (post-edit and Cursor pre-edit) now resolves the project platform via loadContext + extractPlatform and skips its web rule scan for ios / android / adaptive projects, so React Native / Flutter code never draws web-shaped findings (new hook-lib resolveProjectPlatform / isNativePlatform helpers, covered by unit and subprocess tests). - context.mjs CLI warns on an unrecognized ## Platform value (e.g. a toolchain name like `flutter`) instead of silently defaulting to web; extractRegister / extractPlatform now share extractSectionValue. - Removed reference/web.md: nothing loaded it; CLAUDE.md carries the "web has no extra rulebook" explanation. - init.md: skip live-mode config (Step 6) for native platforms; note the per-app PRODUCT.md pattern for repos shipping web + native. - android.md: Material-everywhere apps that also ship on iPhone still owe iOS OS guarantees (safe areas, Reduce Motion, edge-swipe back). - ios.md: reworded a design-time line that framed Dynamic Type as an accessibility check (a11y stays owned by audit.md). - Renumbered the new skill-behavior scenario to 14 after main's 10-13; updated CLAUDE.md scenario list; added android + unrecognized-value CLI test cases. - No version or changelog changes: versioning happens at release time. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Tighten platform reference prose Editorial pass on the platform-axis text, applied with AI assistance (Claude Code) under maintainer direction: - ios.md / android.md rewritten to house style: single-line paragraphs (no hard wraps), one-sentence scope intro, deduplicated intro/slop-test, register-compression down to two sentences. In-file attribution paragraphs removed (NOTICE.md owns attribution); "read on top of the register reference" cruft removed (SKILL step 5 and the context.mjs directive already say it). Bans sections dropped: they restated the rules above them; the two additive items (tab-bar overload, hover-dependent affordances) folded into rules. ~40% smaller each. - Sub-command Platform sections (adapt, audit, animate, layout), SKILL step 5, init.md platform prose, and the context.mjs directive trimmed the same way. Build (prose validators, counts) and both test runners green. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Treat an empty PRODUCT.md section as absent, not the next heading Copilot review catch: extractSectionValue read the next `## ...` heading as the section value when a field was left empty, which made the CLI warn "value `## Product Purpose` is not recognized". Stop at the next heading and return null instead. Regression tests for extractPlatform, extractRegister, and the CLI warning path. Applied with AI assistance (Claude Code) under maintainer direction. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Only read a token list of both native targets as adaptive Bugbot catch: after the exact platform tokens failed, any Platform line containing the words ios and android was classified adaptive, so negated or explanatory prose ("web only, not ios or android") silently loaded both native refs and skipped the hook, with no warning. The combo parse now accepts only list separators and the two platform words; anything else falls through to the CLI's unrecognized-value WARNING. Regression tests added. Applied with AI assistance (Claude Code) under maintainer direction. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Paul Bakaus <paul.bakaus@gmail.com> |
||
|
1a46353b29 |
Don't force init on scoped commands when PRODUCT.md is missing (#277)
* Don't force init on scoped commands when PRODUCT.md is missing Setup step 1 told the agent: "If it reports NO_PRODUCT_MD, stop and follow reference/init.md before doing anything else." For a project with no PRODUCT.md, that turned every scoped request (polish, critique, audit, layout, ...) into a full from-scratch init detour. The user asks to polish one button and the skill instead starts writing PRODUCT.md from the beginning. Faced with that gate, agents also frequently abandon the command and do an ad-hoc pass without loading the command reference. Make the gate command-aware. A missing PRODUCT.md still routes into init for the from-scratch build flows where captured product context is the point (init, craft, shape). For any other command, a scoped request against existing code, the code is the context: proceed with the requested command, infer the register from the surface in focus, and offer /impeccable init once as a suggestion rather than a blocker. - skill/SKILL.src.md: rewrite the step 1 NO_PRODUCT_MD rule; reconcile the no-argument routing rule so it leads the menu with init instead of silently jumping into it; extend the craft init-then-resume footnote to cover shape, now also a from-scratch flow. - skill/scripts/context.mjs: soften the NO_PRODUCT_MD message to defer to the step 1 rule instead of "Stop the current task"; refresh the stale file-level JSDoc that still described the old empty-stdout signal. - tests/skill-behavior/scenarios.test.mjs: add scenario 10 (scoped command + no PRODUCT.md proceeds without forcing init) and scenario 11 (shape + no PRODUCT.md still diverts into init). Scenario 1 (craft diverts) stays green and pins the build path. Source-only per repo convention; provider and plugin copies are regenerated by the maintainer's build:skills sync. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * Fix missing-context routing for build intent --------- Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> Co-authored-by: Paul Bakaus <paul.bakaus@gmail.com> |
||
|
Paul Bakaus
|
6443980117 | Improve CLI install prompts | ||
|
9ffd3211d5 |
Neo Kinpaku design system + Live Mode v3 (#169)
* Add neo kinpaku design system page * skill: rip out baked-in category recipes and saturated-default motion tropes Programmatic bias mining (impeccable-evals) traced four major defects back to specific lines in this skill that contradicted SKILL.md's own first-order-reflex warning: - brand.md "Pairing and voice" prescribed four category→aesthetic recipes (editorial → serif+sans, tech/dev/fintech → tight tracking, consumer/food/travel → script/display serif, creative → rule-break). These directly drove OpenAI's 76% extreme-negative letter-spacing on tech briefs and Anthropic/Google's 28-34% italic-serif-display slop on editorial/food briefs. Replaced with one sentence: the shape depends on the brand, not on the brand's category. - brand.md "Brand permissions" had "Typographic risk. Enormous display type, unexpected italic cuts, mixed cases, hand-drawn headlines, a single oversize word as a hero." — a four-for-one slop driver behind 97% OpenAI comically-large H1, 42% bad-SVG illustration, and the editorial-italic slop. Deleted outright. - typeset.md and teach.md repeated the same category recipes; trimmed to the principle without the recipe. - SKILL.md Typography: added a hard hero-H1 ceiling (clamp() max ≤ 6rem ≈ 96px), with a <codex> block to make it explicit since OpenAI over-indexes here (97% ≥128px vs 24% for Anthropic). - animate.md, bolder.md, brand.md: removed "staggered reveals" and "scroll-triggered transitions" as the prescribed default ambitious motion. By 2026 that's the saturated AI tell, not a choreography. Reserved stagger for legitimate list-sibling rhythm. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * skill: anti-cream + codex-specific defect bans + universal slop bans Second pass after measuring more biases against the eval corpus. - SKILL.md Color: explicit "cream/sand/beige body bg is the saturated AI default of 2026" rule. Tone down the "tint every neutral" line so it doesn't read as "default to warm-tinted near-white" (which OpenAI hits at 74% and Anthropic at 31%-47%). - SKILL.md Absolute bans: add universal bans for two slop patterns detected at 55-95% across providers — tiny uppercase tracked eyebrow above every section (the 2023-era kicker that's now AI grammar) and numbered section markers (01/02/03). Also explicit "text that overflows its container is the universal defect on tablet/mobile." - SKILL.md Absolute bans → <codex> block: ban the GPT-specific defects Paul annotated repeatedly — `border:1px solid` + soft-wide-shadow (≥16px blur) "ghost cards", `border-radius:32px+` over-rounding, hand-drawn/sketchy SVG illustrations (loose-sketch / *-sketch classes, feTurbulence paper-grain filters), repeating-linear-gradient stripes, "X theater" AI-slop copy phrases. - SKILL.md Motion → <gemini> block: the image :hover transform tell (38% Google skill-on rate). Hover effects on images add no info; the image isn't an action target. Animate card chrome, not the image. - SKILL.md Typography: hard display letter-spacing floor ≥-0.04em (OpenAI defaults to -0.075em → cramped). Existing hero ceiling <codex> block extended with the letter-spacing rule. - codex.md Step A example: stop seeding "warm-grounded (deep oxblood + cream)" as the warm-palette template, which primes the cream default. - colorize.md Tinted backgrounds: stop printing the literal cream recipe `oklch(97% 0.01 60)`; replace with brand-anchored guidance. - document.md examples: warm-ash-cream → cool-paper so the example doesn't seed cream as the canonical neutral example. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * skill: universal anti-slop bans + contrast/font-count/all-caps-body rules Third pass after measuring the rest of the cross-provider matrix: - Color: explicit "Verify contrast" rule. Low-contrast text fires at 68% across all providers skill-on (90+% off). The most common failure is muted gray body on a tinted near-white; light-gray-for- elegance is named as the single biggest cause of unreadable AI pages. - Typography: max-3-font-families rule. Overused-fonts (>4 families) fires at 28% Anthropic / 36% Google / 0% OpenAI skill-on; >50% off. Also: universal "no all-caps body copy" (moved from brand-only ban to Shared design laws since product-register also overuses caps). - Copy: anti-aphoristic-cadence ban targets Anthropic's signature "X. No Y." / "X. Just Y." voice (63% skill-on copy-slop rate, 77% off — the worst rate in the matrix). Once-is-voice / three-or-more- is-tell framing per the runner's copy-slop detector. - Copy: anti-SaaS-buzzword-string ban with the literal phrase list the detector watches for (streamline/empower/supercharge, trusted- by-leading, best-in-class/enterprise-grade/cutting-edge, etc). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * skill: strengthen anti-cream rule across full warm-neutral band Smoke validation showed the cream fix worked for Google + OpenAI but Anthropic Sonnet italian-restaurant still shipped `--paper: oklch(90% .018 88)` — cream just outside the L≥95% band the rule cited. Broaden the rule: - Band: OKLCH L 0.84-0.97, C < 0.06, hue 40-100 (was 95-97% / 60-95). - Name the token-name tells explicitly (paper / cream / sand / bone / flour / linen / parchment / wheat / biscuit / ivory) — the model defaults to one of these regardless of what hex it lands on. - Call out the specific brief patterns ("warm, traditional, family- coastal-Italian" / "editorial-restraint") that the model translates into cream by reflex. Then provide three explicit non-cream options: saturated brand color, true off-white at C=0, or darker mid-tone. Warmth in the brand is carried by accent + typography + imagery, not by body bg. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * v3.2.0: skill bias-fix release Bumps version from 3.1.1 to mark the four-commit skill cleanup that rips out baked-in category recipes (brand.md), saturated-default motion tropes (staggered reveals everywhere), the cream/sand body-bg AI tell, codex-specific defects (1px+wide-shadow, over-rounding, hand-drawn SVGs, stripes, X-theater copy), the extreme-letter-spacing default, and universal slop bans (all-caps eyebrow on every section, numbered-section markers, all-caps body, font-family-count > 3, aphoristic copy cadence, SaaS buzzword strings). Plus a hard hero-H1 ceiling (clamp() ≤6rem) and a Gemini-specific image:hover transform block. Validated against ~190 post-fix samples — see impeccable-evals biases tab for per-provider deltas. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * drop "no pure black/white" rule entirely The rule was contested in the design world and causing more damage than good — pushing every page into the tinted-near-white default which is the cream/sand AI tell we already explicitly ban elsewhere. Vercel, SVKMS, Brutalist sites, et al. use pure black/white successfully; the skill shouldn't second-guess that. Skill markdown deletions: - SKILL.md Color: drop the "Never use #000 or #fff" bullet. - color-and-contrast.md: drop the "Never Use Pure Gray or Pure Black" subsection, the "Never pure black" table-row prescription, and the "Avoid: Using pure black for large areas" bullet. - colorize.md: drop the "NEVER use pure black or pure white for large areas" bullet. - polish.md: drop the "Tinted neutrals: No pure gray or pure black" half of the bullet (the gray-on-color bullet survives). Detector code (cli/engine): - registry/antipatterns.mjs: remove the `pure-black-white` entry. - rules/checks.mjs: remove the three `findings.push({ id: 'pure-black-white', ... })` emit points (inline #000 bg, Tailwind bg-black class, plain-HTML scan path). - engines/regex/detect-text.mjs: remove the two pure-black-white regex rules (CSS `background: #000…` + Tailwind `bg-black`). - detect-antipatterns-browser.js: regenerated via scripts/build-browser-detector.js. Tests: - detect-antipatterns-fixtures.test.mjs: invert the assertion that pure-black-white fires; expect it to NOT fire post-v3.2. Drop the Tailwind bg-black-opacity edge-case test (no longer relevant). - detect-antipatterns.test.js: drop the standalone "detects pure- black-white in styled-components" test and remove pure-black-white from the multi-detector assertions in PricingCard, globals.css, and GlobalStyle.tsx tests. 166 bun tests pass; 24 node fixture tests pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * skill: strip example patterns from copy rules, strengthen gemini block v3.2 rerun validation surfaced two issues: 1. Copy-slop detector fires more on Gemini under v3.2 (48% → 84%) than under no-skill baseline. Root cause: the anti-aphoristic-cadence rule printed the literal "X. No Y." / "X. Just Y." patterns as examples, and Gemini imitated them as the recommended voice. Same recipe-becomes- bias trap we hit with brand.md:116's "Enormous display type, unexpected italic cuts, mixed cases, hand-drawn headlines" enumeration. Fix: describe the cadence as a rhythm ("serious statement, then punchy short negation") without printing literal patterns. Buzzword list trimmed to a single inline phrase family rather than quoted strings. 2. Gemini image:hover transform Gemini-tell hadn't dropped (31% off → 32% v3.2). Strengthen the <gemini> block: explicit "Never animate <img> elements on hover", call out the Tailwind group-hover:scale / group-hover:rotate / group-hover:translate parent-hover patterns by name (Gemini was reaching for these via Tailwind even though the prior text talked about :hover on the image directly). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * skill: simplify context loading and inline register directive Replaces load-context.mjs's JSON output with a tight markdown block from the renamed context.mjs. The script now extracts PRODUCT.md's `## Register` field and appends a `NEXT STEP:` directive naming the matching reference (brand.md / product.md), which moved Gemini from skipping the register load entirely to honoring it. Drops the `.impeccable.md` auto-migration; makes IMPECCABLE_CONTEXT_DIR a lazy escape hatch consulted only when the default paths come up empty. Setup is now four bullets in one list. The DESIGN.md nudge is gone; in its place, a "familiarize with the existing design system" step that calls out CSS / tokens / running app as authoritative sources alongside DESIGN.md. The standalone `### Register` H3 stays for the cascade rules (task cue → surface → register field). New LLM-backed test suite at tests/skill-behavior/ runs five scenarios against claude-haiku-4-5, gpt-5.4-mini, and gemini-3.1-flash-lite via Vercel AI SDK. Captures real tool traces, asserts on context.mjs calls, brand.md loads, and teach.md fallback. Skips cleanly when API keys are unset. 13-14/15 pass; only stable failure is the v3.2.0-era gpt-mini S4 "don't re-run" regression. Adds @ai-sdk/google as devDep and the test:skill-behavior npm script. Touches em-dashes in skill/SKILL.md and four reference files so `bun run build:skills` passes its skill-prose validator. teach.md and document.md drop their "re-run the loader to refresh session cache" steps since the agent's own write is now the freshest source. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * skill: merge orphan reference files into command sub-skills + inline S-tier invariants Two related restructurings: 1. SKILL.md now carries the cross-domain invariants that catch defects in any project (contrast/placeholder/gray-on-color, similar-font pairing, text-wrap, tabular-nums, centered-stack default, Flex/Grid choice, auto-fit grids, semantic z-index, reduced motion, stagger vs section-fade, premium motion materials, focus-visible, placeholders-aren't-labels, dropdown overflow trap, button/link copy). Greenfield-only rules (theme picking, color strategy, tinted neutrals) live under "New projects only". 2. Reference files merged into their command counterparts: - spatial-design.md -> layout.md - motion-design.md -> animate.md - color-and-contrast.md -> colorize.md - responsive-design.md -> adapt.md - ux-writing.md -> clarify.md - typography.md -> typeset.md (bolder.md redirected) - cognitive-load.md + heuristics-scoring.md + personas.md -> critique.md craft.md and shape.md "load references" lists updated to new file homes. interaction-design.md stays standalone (no 1:1 command verb). Net: 36 -> 27 reference files. Same content, fewer files, no orphaned reference loaded only from craft.md. Also extends the routing rules: if the user's first word doesn't match a command but the intent clearly maps to one, load that command's reference and proceed as if invoked. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * skill: add sub-command + existing-project scenarios; move sub-command load to step 2 Adds three new LLM-backed scenarios to tests/skill-behavior: - S6: `/impeccable polish` → loads polish.md - S7: `/impeccable audit` → loads audit.md - S8: existing SvelteKit project (PRODUCT.md + DESIGN.md + src/app.css + src/lib/components/*.svelte + src/routes/+page.svelte) → agent reads at least one project code file to understand the existing design system S6/S7 surface a real model-floor: gpt-5.4-mini reads brand.md, reads the target index.html, and just does the polish/audit without ever loading the sub-command reference. Stronger SKILL.md wording didn't move it. Captured in the README baseline as a known weakness. Claude and Gemini honor the load reliably. To fix Gemini on S6/S7, sub-command reference loading is now Setup step 2 (right after context.mjs), not step 4 — placing it before the model gets focused on "doing the work". Step 3 (design-system familiarization) is tightened to require at least one project code read even when a sub-command reference loads in step 2, so Claude doesn't laser-focus on the sub-command flow and skip the broader exploration. Two new fixtures: MINIMAL_LANDING_HTML (a tiny static landing page for S6/S7) and SVELTE_PROJECT_FILES (a minimal SvelteKit scaffold with tokens, components, and a routes/+page.svelte for S8). Both designed to look real enough that agents treat them as production code. Suite is now 24 tests across three providers; baseline is 21-22/24, with the stable failures being gpt-5.4-mini scenarios 6 and 7. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * skill: add reveal-animation safety rule (must enhance, not gate visibility) Class-triggered visibility transitions pause on hidden tabs and headless renderers. The italian-restaurant smoke produced a build where 2 sections shipped opacity:0 because the CSS transition never advanced past currentTime=0 (timeline paused). Added one-liner under Motion to prevent the antipattern: reveals must enhance an already-visible default, never gate content visibility on a class-triggered transition. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * skill: restore prescriptive cream/sand/beige paragraph Bisection across 5 historical skill commits on Gemini 3.5 flash fast lane n=3 found that |